Github user revans2 commented on the pull request:
https://github.com/apache/storm/pull/797#issuecomment-148393429
@darionyaphet I have seen only small changes to the performance of storm.
It is close to nothing really, but in my micro-benchmarks I see about between a
6% and 26% improvement in maximum throughput.
The big reason for upgrading is that the APIs are improved so that we can
support #765. In that I have seen a maximum throughput for a single node more
then double. Making a faster version of wordcount go from 10,000
sentences/second fully counted to 22,000, on my Mac Book Pro laptop. If you
include acking and everything else that happens that is going from 320,000
tuples/second sent+processed to 704,000 tuples/second. Be aware that you are
not likely to see this in real life on a single topology, as the messaging
layer is likely to slow things down, and wordcount is a very light weight
processing. Real world processing is usually more intensive.
But this is just the beginning we are working on benchmarks and subsequent
optimizations all throughput the critical path of storm.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
